W.Gen {LPKsample}

Similarity matrix based on eLP basis and polynomial kernel

Description

Given data matrix X and eLP order k, this function generate the similarity matrix W for graph analysis.

Usage

W.Gen(X, k, c.poly = 0.5)

Arguments

X

A n-by-d matrix of the observations

k

An integer, order of LP component

c.poly

Numeric, parameter for polynomial kernel

Value

A n-by-n similarity matrix generated from k-th order eLP transformation of X

Author(s)

Mukhopadhyay, S. and Wang, K.

References

Mukhopadhyay, S. and Wang, K. (2020), "A Nonparametric Approach to High-dimensional K-sample Comparison Problem", arXiv:1810.01724.

See Also

GLP

Examples

#example: 6 observations on 3 features:
 x<-rbind(matrix(runif(9),3,3),matrix(runif(9)+1,3,3))
#LP similarity matrix:
 simmat<-W.Gen(x,1)$W
 image(simmat)